ZONOS2 Voice Cloning TTS
Short-reference narration, voice tests, and multilingual speech samples
Input
Reference voice audio plus multilingual text for ZONOS2 voice cloning TTS
Workflow
- Upload a clean reference voice clip in a supported audio format
- Enter text, select its language and adjust settings as needed
- Generate speech and download the 44.1 kHz WAV result
Parameter guide
- Reference audio: uploaded speaker clip used for zero-shot voice cloning.
- Text to synthesize. The selected language controls normalization.
- Language: official text-normalization code such as English (US), Chinese (Mandarin), Japanese, or Korean.
- Quality settings: accurate mode, quality conditioning and the clean-reference setting are enabled by default.
- Seed: optional deterministic seed for repeatable comparison runs.
What to check
- Use a clear reference voice with minimal background noise